Mechanism and application of a microcapsule enabled multicatalyst reaction

Article abstract of DOI:10.1021/ja071706x

In this paper, we describe the development and application of a multistep one-pot reaction that is made possible by the site isolation of two otherwise incompatible catalysts. We prepared a microencapsulated amine catalyst by interfacial polymerization and used it in conjunction with a nickel-based catalyst for the transformation of an aldehyde to a Michael adduct via a nitroalkene intermediate. The amine-catalyzed conversion of an aldehyde to a nitroalkene was found to proceed through an imine rather than a nitroalcohol. Kinetic studies indicated that the reaction is first order in both the nickel catalyst and the shell of the encapsulated amine catalyst. Furthermore, we provide evidence against interaction between amine and nickel catalysts and present kinetic data that demonstrates that there is a rate enhancement of the Michael addition due to the urea groups on the surface of the microencapsulated catalyst. We applied our one-pot reaction to the development of a new synthetic route for pregabalin that proceeds with an overall yield of 74%.

Full text of DOI:10.1021/ja071706x

Published on Web 06/30/2007
Mechanism and Application of a Microcapsule Enabled
Multicatalyst Reaction
Sarah L. Poe, Muris Koba sˇ lija, and D. Tyler McQuade*
Contribution from the Department of Chemistry and Chemical Biology, Baker Laboratory,
Cornell UniVersity, Ithaca, New York 14853
Received March 11, 2007; E-mail: dtm25@cornell.edu
Abstract: In this paper, we describe the development and application of a multistep one-pot reaction that
is made possible by the site isolation of two otherwise incompatible catalysts. We prepared a microen-
capsulated amine catalyst by interfacial polymerization and used it in conjunction with a nickel-based catalyst
for the transformation of an aldehyde to a Michael adduct via a nitroalkene intermediate. The amine-catalyzed
conversion of an aldehyde to a nitroalkene was found to proceed through an imine rather than a nitroalcohol.
Kinetic studies indicated that the reaction is first order in both the nickel catalyst and the shell of the
encapsulated amine catalyst. Furthermore, we provide evidence against interaction between amine and
nickel catalysts and present kinetic data that demonstrates that there is a rate enhancement of the Michael
addition due to the urea groups on the surface of the microencapsulated catalyst. We applied our one-pot
reaction to the development of a new synthetic route for pregabalin that proceeds with an overall yield of
74%.
Introduction
sophisticated systems, the full potential of multicatalyst systems
will remain unrealized until applications to more complex
molecules are demonstrated.
Catalyst isolation techniques that enable one-pot multistep
reactions hold great potential for increasing the efficiency of
chemical synthesis. Performing multiple reactions simulta-
neously in a single reaction vessel offers possibilities for reduced
With this need in mind, we recently reported a microcapsule
(
µcap) enabled multicatalyst system that produces a synthetically
7
1
useful product. The reaction involves the amine-catalyzed
transformation of an aldehyde to a nitroalkene followed by a
transition-metal-catalyzed Michael addition in the same reaction
vessel (Scheme 1). Typically, amine catalysts and nickel
complexes are incompatible due to their tendency to chelate
waste and increased safety as well as manipulation of equilibria.
Although many site-isolated catalysts have been developed, the
focus has been largely on catalyst recovery rather than on
tandem catalysis. Indeed, since the pioneering work of Patchornik
over 25 years ago the sol-gel materials developed by Avnir
and Blum have been the only catalyst supports used in the
context of one-pot multistep catalysis until recently. In the past
few years, Fr e´ chet et al. introduced a multicatalyst system that
employed star polymers for catalyst isolation, while Jones et
2
7
and render each other inactive. However, microencapsulation
3
of PEI forms catalyst 1, which can successfully be used in
tandem with the nickel-based catalyst 2 developed by the Evans
8
4
group.
Not only do these two reactions both form C-C bonds, but
together they create a versatile synthetic building block. For
instance, the nitroalkane can be converted into an amine via
reduction or a carbonyl via the Nef reaction, while the ester
groups can be transformed into a single carboxylate via
hydrolysis-decarboxylation or a diol via reduction. Such
subsequent reactions could provide access to a wide range of
useful intermediates, such as those found en route to pharma-
ceuticals ranging from rolipram to baclofen to pregabalin (3,
Scheme 2).
In this article, we show the generality of this reaction while
providing mechanistic insight for our microencapsulated catalyst.
We also discuss the application of this system to a pharmaceuti-
cally relevant problem: an efficient chemical synthesis of
pregabalin.
al. showed that catalysts supported on polymer resins and
magnetic particles could be used together while avoiding catalyst
fouling. In another approach Kaneda et al. immobilized two
5
incompatible catalysts in mesoporous clays to achieve site
6
isolation. All of these examples demonstrate the effective
separation of otherwise incompatible catalysts. Though each of
these examples promises the possibility for application to more
(
1) (a) Hall, N. Science 1994, 266, 32-34. (b) Tietze, L. F. Chem. ReV. 1996,
9
6, 115-136. (c) Wasilke, J.-C.; Obrey, S. J.; Baker, R. T.; Bazan, G. C.
Chem. ReV. 2005, 105, 1001-1020.
(
2) Cohen, B. J.; Kraus, M. J.; Patchornik, A. J. Am. Chem. Soc. 1981, 103,
7
620-7629.
(
3) (a) Gelman, F.; Blum, J.; Avnir, D. J. Am. Chem. Soc. 2000, 122, 11999-
12000. (b) Gelman, F.; Blum, J.; Avnir, D. Angew. Chem., Int. Ed. 2001,
40, 3647-3649. (c) Gelman, F.; Blum, J.; Avnir, D. New J. Chem. 2003,
27, 205-207.
(
(
(
4) Helms, B.; Guillaudeu, S. J.; Xie, Y.; McMurdo, M.; Hawker, C. J.; Fr e´ chet,
J. M. J. Angew. Chem., Int. Ed. 2005, 44, 3684-6387.
5) Phan, N. T. S.; Gil, C. S.; Nguyen, J. V.; Zhang, Z. J.; Jones, C. W. Angew.
Chem., Int. Ed. 2006, 45, 2209-2212.
(7) Poe, S. L.; Koba sˇ lija, M.; McQuade, D. T. J. Am. Chem. Soc. 2006, 128,
15586-15587.
(8) Evans, D. A.; Seidel, D. J. Am. Chem. Soc. 2005, 127, 9958-9959.
6) Motokura, K.; Fujita, N.; Mori, K.; Mizugaki, T.; Ebitani, K.; Kaneda, K.
J. Am. Chem. Soc. 2005, 127, 9674-9675.
9216
9
J. AM. CHEM. SOC. 2007, 129, 9216-9221
10.1021/ja071706x CCC: $37.00 © 2007 American Chemical Society

Microcapsule Enabled Muticatalyst Reaction
A R T I C L E S
Scheme 1 Transformation of an Aldehyde to a Nitroalkene and
Subsequent Michael Addition of a Malonate Ester Can Be
Performed in Tandem through the Use of Site-Isolated Catalystsa
a
The two catalysts are microencapsulated PEI (1) and a nickel-based
complex (2).
Scheme 2. Pharmaceutical Agents That Incorporate an Aldehyde,
a Nitroalkane, and a Malonate Ester
Figure 1. Optical micrographs of dry µcaps (a); µcaps in methanol, a
swelling solvent (b); and µcaps in toluene, a nonswelling solvent (c). The
scale bar is 30 µm.
Results and Discussion
Synthesis and Characterization of Microencapsulated
Catalyst. Encapsulation of an amine-based Henry reaction
catalyst was achieved via the interfacial polymerization of oil-
in-oil emulsions, as described in our previous work. Poly-
(ethyleneimine) (PEI) was encapsulated by dispersing a meth-
anolic PEI solution into a continuous cyclohexane phase. Upon
emulsification, 2,4-tolylene diisocyanate (TDI) was added to
initiate cross-linking at the emulsion interface, forming polyurea
shells that contain free chains of PEI. The microcapsules crenate
when dry and swell when placed in such solvents as methanol
and DMF, suggesting a hollow capsule rather than a solid sphere
9
(
g
Figure 1). Catalyst loading was determined to be 1.6 mmol/
1
0
by acylation of the catalytic amines with trifluoroacetic
anhydride followed by fluorine elemental analysis. Oxygen
elemental analysis placed the upper limit on urea content of
the µcap shells at 5.6 mmol/g (the upper limit assumes no water
retention).
Figure 2. Conversion of benzaldehyde (4) after 6 h for the amine-catalyzed
reaction between benzaldehyde and nitromethane. Catalysts for the reaction
were free PEI (black bars, 4.6 mol %) and encapsulated PEI (white bars,
Activity and Mechanism of Microencapsulated Catalyst.
To better understand the importance of µcap swelling, the
reaction between benzaldehyde (4) and nitromethane was
performed in a range of different solvents.11 Swelling effects
4.6 mol %).
encapsulated PEI as catalysts for formation of trans-â-nitrosty-
rene (5) and 1,3-dinitro-2-phenyl-propane (6). Figure 2 shows
benzaldehyde conversion after 6 h for each catalyst.
were separated from solvent effects by using both free and
(
9) For a detailed characterization of oil-in-oil microcapsules, see: Koba sˇ lija,
The results for the reactions catalyzed by free PEI (black bars)
are not affected by swelling or by the kinetic barrier introduced
by the microcapsule shell and therefore indicate how effective
M.; McQuade, D. T. Macromolecules 2006, 39, 6371-6375.
(
(
10) We define catalytic amines as the primary amines of PEI.
11) See Supporting Information for full range of solvents.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 29, 2007 9217

A R T I C L E S
Poe et al.
each solvent is for this reaction. If the differences in conversion
for the reactions catalyzed by encapsulated PEI (white bars)
were based exclusively on solvent, we would expect the two
cases to show the same trends. This is the case for both swelling
and partially swelling solvents; for each catalyst conversions
are high for ethanol, moderate for chloroform, and low for
acetone. However, this is not true for nonswelling solvents.
While the free PEI-catalyzed reactions revealed that toluene,
ether, and THF were relatively good, moderate, and poor
solvents, respectively, encapsulated PEI did not produce the
same results. Despite the moderate conversion produced by free
PEI in toluene, encapsulation resulted in an 80% decrease in
catalytic activity in the same solvent. These results suggest that
the solvent dependence of this reaction is 2-fold; not only must
the solvent be favorable for the PEI-catalyzed reaction, but it
also must be able to swell the microcapsules. Acetone swells
the capsules but is a poor solvent for the reaction, while toluene
is a good solvent for the reaction but is unable to swell the
µcaps. Both of these cases result in poor conversions of
benzaldehyde when encapsulated PEI is used as the catalyst.
Only ethanol, a good solvent for the reaction that is also able
to swell the capsules, is able to produce high conversions with
both free and encapsulated PEI. Furthermore, in some cases
catalytic activity is retained for capsules that are swollen in a
swelling solvent and then placed in a bulk nonswelling solvent
Scheme 3. Proposed Catalytic System of µcap-Catalyzed
Nitroalkene Formationa
a
Though the mechanism for nitroalkene formation and regeneration of
the primary amine (left) is believed to occur intermolecularly, it is shown
intramolecularly for clarity.
(
Figure 2). We are currently pursuing the reasons for and
12
implications of this phenomenon.
Scheme 4. Single-Catalyst Addition of Nitromethane (top) versus
Double-Catalyst Addition of Dimethyl Malonate (DMM) (bottom)
With a better understanding of the conditions required for
catalytic activity, we turned our attention to the mechanism of
the µcap catalyst. Transformation of an aldehyde to a nitroalkene
can occur via two different pathways. The first involves
nitroalcohol formation through a traditional Henry reaction,
which is then followed by elimination to form the double bond.
The second proceeds through an imine intermediate rather than
the nitroalcohol. This latter mechanism has been suggested for
cases in which the catalyst contains both primary and tertiary
1
3
amino groups as well as for solid supported catalysis. Being
that our catalyst exhibits both of these features, we predicted
that µcap-enabled nitroalkene formation goes through an imine
intermediate rather than the nitroalcohol. Indeed, when we
followed the µcap-catalyzed condensation of benzaldehyde with
nitromethane over time, we did not observe the nitroalcohol at
any point in the reaction. However, the nitroalcohol was found
to be present during the course of the one-pot reaction, possibly
having been formed by the amine ligands of the nickel catalyst.
This evidence precludes the possibility that elimination occurs
as soon as the nitroalcohol is formed and suggests that the
reaction might follow the latter route. To further support this
hypothesis, when the nitroalcohol is placed in the presence of
swollen µcaps, no nitroalkene formation is observed. The
inability of the µcap catalyst to convert this potential intermedi-
ate to the final product provides further evidence against the
nitroalcohol-elimination pathway. A proposed mechanism for
µcap-catalyzed nitroalkene formation is shown in Scheme 3.
One-Pot Reaction. We reported that there is the possibility
for the nitroalkene intermediate to either form the dinitro product
or go through a Michael-type addition with the encapsulated
PEI when subjected to the reaction discussed above. However,
we have also shown that when the reaction is run in the presence
of a second catalyst-reagent pair, this intermediate can be
trapped and directed to a different reaction pathway (Scheme
4
).
Furthermore, because µcaps swollen in methanol retain their
catalytic activity when placed in toluene, the reaction can be
run in a mixture of two different solvents. This allows both the
µcaps and the nickel catalyst to operate in their respective ideal
solvents of methanol and toluene. To demonstrate the scope of
this one-pot reaction, we performed this reaction with a variety
of aromatic and aliphatic aldehydes. The results are shown in
Table 1. It is evident that though the system tolerates both
aromatic and aliphatic aldehydes, introduction of electron-
withdrawing substituents on the aromatic substrates results in
decreased yields. This effect is most pronounced in entries 6
and 7 for which the cyano- and nitro-substituted benzaldehydes
yield minimal product formation. Assuming that the µcap-
catalyzed step proceeds through an imine intermediate, this
phenomemon can be rationalized by the observations of Santerre
et al., who compared the rates of uncatalyzed imine formation
(
12) Koba sˇ lija et al. Unpublished results.
(
13) (a) Tamura, R.; Sato, M.; Oda, D. J. Org. Chem. 1986, 51, 4368-4375.
(b) Demicheli, G.; Maggi, R.; Mazzacani, A.; Righi, P.; Sartori, G.; Bigi,
F. Tetrahedron Lett. 2001, 42, 2401-2403.
9
218 J. AM. CHEM. SOC. VOL. 129, NO. 29, 2007
9

Microcapsule Enabled Muticatalyst Reaction
A R T I C L E S
Figure 3. Kinetic studies on the tandem reaction of 3-methylbutyraldehyde,
nitromethane, and dimethyl malonate.
Table 1. Scope of the One-Pot Reaction
entry
R
product
yield (%)
1
2
3
4
5
6
7
8
9
Ph
7a
7b
7c
7d
7e
7f
7g
7h
7i
80
94
89
43
48
<5
<5
71
65
4-Me-Ph
4-MeO-Ph
4-Br-Ph
4-Cl-Ph
4-CN-Ph
4-NO2-Ph
CH(Me)2
CH2CH(Me)2
Figure 4. µcap-accelerated Michael addition between benzaldehyde (4)
and dimethyl malonate in the presence of untreated µcaps (a) and in the
presence of acylated µcaps (b).
between substituted benzaldehydes and aliphatic amines at room
1
4
temperature. It was found that the rate is maximal for
unsubstituted benzaldehyde and steadily decreases as the
substituents’ σ values diverge in either direction on the Hammett
plot. It is possible that the cyano and nitro substituents are too
electron withdrawing to allow for any appreciable imine
formation.
therefore instructive to determine whether the rate-determining
step is at all retarded by interaction between the two catalysts.
We began our investigation by comparing the rates of the
Michael addition in the absence and presence of the microen-
capsulated catalyst. One complicating factor, however, is that
the catalytic amines within the µcaps irreversibly react with
nitroalkenes, decreasing the amount of starting material available
for reaction. Using the Michael addition between trans-â-
nitrostyrene (5) and dimethyl malonate as a model, we ap-
proached this problem in two ways. The first approach was to
To gain information about the mechanism of the overall
tandem reaction, we carried out kinetic studies to identify the
rate-determining step. Changing the catalyst concentration in
the reaction between 3-methylbutyraldehyde (8), nitromethane,
and dimethyl malonate revealed that the reaction is first order
in nickel catalyst 2 (Figure 3), indicating that the Michael
addition of dimethyl malonate to the nitroalkene is the rate-
determining step.
“normalize” the data from the µcap-containing reaction in order
to account for the loss of starting material (Figure 4a). Product
yields were calculated using the formula (mol of 7a)/(mol of 5
+
mol of 7a). It should be noted that these calculations correct
The kinetic data shown in Figure 3 reveal that the rate of the
tandem reaction depends on the concentration of nickel catalyst
and consequently on the efficiency of site isolation. It is
only for decreased product formation due to nitroalkene-µcap
interaction; any product suppression due to catalyst interaction
should still be apparent. Both reactions attain an adjusted yield
of 90% after 10 h, indicating that the presence of the µcaps
does not depress the rate of the Michael addition. Indeed, they
2
(
14) Santerre, G. M.; Hansrote, C. J.; Crowell, T. L. J. Am. Chem. Soc. 1958,
8
0, 1254-1257.
J. AM. CHEM. SOC.
9
VOL. 129, NO. 29, 2007 9219

A R T I C L E S
Poe et al.
Scheme 5. Synthesis of Pregabalin (3)a
a
Reagents and conditions: (a) nitromethane, dimethyl malonate, 1, 2,
toluene, methanol, room temperature, 48 h, 94%, 72% ee. (b) Raney Ni,
H2 (45 psi), EtOH, room temperature, 18 h, 96%. (c) 5 M HCl, 115 °C, 18
h, 95%, 72% ee.
Figure 5. Order plot for the Michael addition between benzaldehyde (4)
and dimethyl malonate in the presence of acylated µcaps. Rate is plotted
as a function of nickel catalyst 2. Mole % urea is based on oxygen content
determined by elemental analysis.
enantioselectively. The desirability of performing these reactions
in tandem rather than sequentially is evident when one considers
the difficulty of isolating of the nitroalkene intermediate. An
efficient nitroalkene-forming reaction between 3-methylbutyr-
aldehyde (8) and nitromethane has only been reported once16
and in our hands often yields a significant amount of byproducts.
However, we demonstrated that with our one-pot reaction there
is no need for nitroalkene isolation and the byproducts are
avoided.
The total synthesis of pregabalin is depicted in Scheme 5.
Using 3-methylbutyraldehyde (8) as the starting material and
an enantioselective version of nickel catalyst 2, the tandem
reaction produced the corresponding Michael adduct (S)-7i in
Figure 6. Proposed transition state for the tandem reaction.
9
4% yield and 72% ee. It should be noted that this tandem
instead appear to provide rate enhancement as the µcap-
containing reaction reached its final yield after only 4 h.
To demonstrate that the apparent rate enhancement is not
merely an artifact of data correction, we acylated the reactive
amines in order to prevent their reaction with 5. The data shown
in Figure 4b confirms the rate enhancement in the presence of
microcapsules. This result is consistent with previous studies
reporting the acceleration of Michael additions by ureas and
thioureas. Since the capsule walls are composed of polyurea,
it is not surprising that we observe the same phenomenon in
our system. Upon further investigation into this rate enhance-
ment this reaction proved to be first order in the concentration
of urea groups (Figure 5). The proposed transition state for the
tandem reaction is illustrated in Figure 6. This finding indicates
that the presence of the urea-containing µcaps accelerates the
Michael addition and ultimately the overall one-pot reaction.
In addition, this rate enhancement does not appear to be
accompanied by any degradation in yield, suggesting effective
site isolation of the two catalysts.
Enantioselective Synthesis of Pregabalin. An attractive
feature of our two-step one-pot reaction is that it not only
incorporates an innovative technique for site-isolation but also
produces synthetically useful products when an enantioselective
version of 2 is used. The Michael adducts that are created are
precursors to γ-amino acids, allowing access to γ-amino butyric
acid (GABA) analogs. Pregabalin is one such analog that is
approved for the treatment of both epilepsy and neuropathic
pain. We imagined a synthesis of pregabalin where our two-
catalyst system forms the pregabalin backbone efficiently and
catalysis system efficiently suppressed the yield of the undesired
dinitro byproduct to less than 5%. Overnight hydrogenation of
(S)-7i with Raney Ni gave nearly quantitative conversion to the
ring-closed product 9. Subsequent acid hydrolysis and decar-
boxylation proceeded in 95% yield of the HCl salt of pregabalin
(3), which retains an ee of 72%. It has been reported that
enantiomeric enrichment of pregabalin can be achieved by
recrystallization for cases in which the ee is at least 85% of the
1
5
1
7
S-enantiomer. Treatment of the 72% ee HCl salt with base
followed by a single recrystallization from isopropyl alchol/
water afforded a product with 91.5% ee. Our successful
enrichment demonstrates the viability of obtaining an enantio-
merically pure product without the need for a resolution step,
which would bring the overall yield of this synthesis to 74% of
enantiomerically pure material.
Conclusion
We developed and evaluated a two-step reaction that is
capable of converting both aromatic and aliphatic aldehydes to
their corresponding Michael adducts in a single reaction vessel.
The two-catalyst system was made possible by microencapsu-
lation of PEI in order to provide site isolation of two otherwise
incompatible catalysts. Kinetic studies indicated that the pres-
ence of the microcapsule shell of catalyst 1 enhanced the rate
of the reaction promoted by catalyst 2. Therefore, the micro-
capsules not only make the reaction possible by providing
catalyst site isolation, but they also serve a second function by
(
16) Kantam, M. L.; Sreekanth, P. Catal. Lett. 1999, 57, 227-231.
17) Bao, J.; Beylin, V. G.; Greene, D.; Hoge, G. S.; Kissel, W.; Marlatt, M.
E.; Pflum, D. A.; Wu, H.-P. U.S. Patent Application 2005/0228190 A1,
2005.
(
(
15) Okino, T.; Hoashi, Y.; Furukawa, T.; Xu, X.; Takemoto, Y. J. Am. Chem.
Soc. 2005, 127, 119-125.
9220 J. AM. CHEM. SOC.
9
VOL. 129, NO. 29, 2007

Microcapsule Enabled Muticatalyst Reaction
A R T I C L E S
enhancing the rate-determining step. On the basis of our
findings, we proposed mechanisms for each step that are
consistent with previously published literature. Using this
tandem reaction we achieved a three-step total synthesis of
pregabalin in 74% overall yield, demonstrating the applicability
of this system to creating synthetically interesting compounds.
By obtaining a better understanding of this complex multicom-
ponent system, we continue to make progress toward the
development of similar tandem reactions.
Acknowledgment. We thank ARO (MAP-MURI), NSF
(SENSORS), NSF (SGER), Nanobiotechnology Center, Tri-
Institutional Program in Chemical Biology, and NYSTAR.
Supporting Information Available: Experimental methods,
catalyst preparation and characterization, and kinetic data. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA071706X
J. AM. CHEM. SOC.
9
VOL. 129, NO. 29, 2007 9221